SambaNova has become a $11 billion company by solving a problem that's quietly reshaping enterprise AI: how to run AI models fast and cheaply once they're built. The company just completed a $1 billion funding round led by General Atlantic, signaling that investors now see inference, not training, as the trillion-dollar opportunity in artificial intelligence.

This shift matters because for years, the AI conversation centered on building bigger models. Companies like OpenAI, Google, and Meta spent billions training massive language models (LLMs), which are AI systems trained on enormous amounts of text data. But once those models exist, the real cost and complexity comes from actually using them. That's where SambaNova enters the picture.

Why Is Inference Becoming More Important Than Training?

Think of it this way: training an AI model is like building a car factory. Inference is like actually manufacturing and selling millions of cars. The factory is expensive to build once, but the ongoing production costs determine whether the business survives.

SambaNova's latest chip, called the SN50, delivers what the company claims is five times faster performance than competing chips while cutting inference costs to one-third of what GPU-based systems cost. The company unveiled the SN50 in February 2026 and announced it would begin shipping to customers later that year. SoftBank Corp. became the first major customer, planning to deploy the chip in its next-generation AI data centers across Japan to serve enterprise and sovereign AI customers throughout Asia-Pacific.

The timing reflects a broader market reality: enterprises are moving past AI experiments and pilots. They're now deploying AI agents, which are autonomous systems that can perform tasks like customer service, code generation, and financial analysis without constant human intervention. These agents need to respond instantly and run continuously, which means inference efficiency becomes a competitive advantage.

"AI is no longer a contest to build the biggest model. With the SN50 and our deep collaboration with Intel, the real race is about who can light up entire data centers with AI agents that answer instantly, never stall, and do it at a cost that turns AI from an experiment into the most profitable engine in the cloud," said Rodrigo Liang, co-founder and CEO of SambaNova.

Rodrigo Liang, Co-founder and CEO of SambaNova

How Does SambaNova's Technology Work Differently?

SambaNova's advantage rests on its Reconfigurable Data Unit (RDU) architecture, a custom-built chip design optimized specifically for running AI models rather than general computing tasks. Unlike graphics processing units (GPUs), which were originally designed for video game rendering and later adapted for AI, the RDU is purpose-built from the ground up for inference workloads.

The SN50 chip includes several technical improvements that translate to real-world benefits for enterprises:

• Ultra-Low Latency: The chip delivers near-instant responses, critical for applications like voice assistants and real-time customer interactions where delays frustrate users.
• Massive Concurrency: It can power thousands of simultaneous AI sessions with consistent performance, allowing a single data center to serve many customers at once without slowdowns.
• Larger Model Support: The three-tier memory architecture can handle models with over 10 trillion parameters and context windows exceeding 10 million tokens, enabling deeper reasoning and richer outputs.
• Cost Efficiency: Higher hardware utilization lowers the cost per token, the fundamental unit of AI processing, driving better return on investment for enterprises.

JPMorganChase, one of the world's largest financial institutions, recently selected SambaNova as an inference infrastructure partner, deploying both the SN40 and SN50 systems for secure, on-premises AI inference. This endorsement from a major enterprise signals confidence in the technology's reliability and performance for demanding workloads.

What Does the Intel Partnership Mean?

SambaNova's collaboration with Intel, announced alongside the SN50 launch, represents a significant bet on heterogeneous AI infrastructure, meaning data centers that combine multiple types of processors rather than relying solely on GPUs. Intel plans to integrate SambaNova's chips with its own processors, networking technology, and storage systems to create a complete alternative to GPU-centric AI deployments.

"Customers are asking for more choice and more efficient ways to scale AI. By combining Intel's leadership in compute, networking, and memory with SambaNova's full-stack AI systems and inference cloud platform, we are delivering a compelling option for organizations looking for GPU alternatives to deploy advanced AI at scale," said Kevork Kechichian, EVP and General Manager of Intel's Data Center Group.

Kevork Kechichian, EVP and General Manager, Data Center Group, Intel

The partnership spans three areas: expanding SambaNova's cloud platform using Intel infrastructure, integrating SambaNova systems with Intel's CPUs and accelerators, and joint sales and marketing to reach enterprises and cloud providers worldwide. Together, the companies aim to unlock what they describe as a multi-billion-dollar inference market opportunity.

How Are Investors Viewing the Inference Market?

The funding round tells a clear story about investor confidence. General Atlantic led the $1 billion Series F round, with participation from major institutional investors including Seligman Ventures, T. Rowe Price Associates, Capital Group, and BlackRock. This follows an earlier $350 million Series E round completed in February 2026, bringing total recent funding to over $1.35 billion.

"SambaNova's platform is differentiated, built for a market where inference has become foundational to enterprise and industry transformation. Rodrigo and the team are driving deep technical innovation to achieve growing commercial momentum while demand for inference is accelerating well ahead of supply," said Martín Escobari, Co-President and Head of Global Growth Equity at General Atlantic.

Martín Escobari, Co-President and Head of Global Growth Equity at General Atlantic

The $11 billion valuation reflects what investors see as a structural shift in AI economics. As enterprises move from experimenting with AI to deploying it in production, the cost of running models becomes as important as the quality of the models themselves. SambaNova's three-times-lower total cost of ownership compared to GPU-based systems directly addresses this concern.

Steps to Understanding SambaNova's Market Position

• Recognize the Shift: The AI industry is transitioning from a focus on training the largest models to optimizing how those models run in production, where inference costs determine profitability.
• Understand the Competition: SambaNova competes not just with GPU makers like Nvidia, but with other custom chip designers and cloud providers building inference-optimized infrastructure.
• Track Customer Adoption: Watch for announcements from major enterprises and cloud providers deploying SambaNova systems, as customer wins validate the technology's real-world effectiveness.
• Monitor the Intel Partnership: The success of SambaNova and Intel's collaboration will determine whether heterogeneous AI infrastructure becomes a viable alternative to GPU-dominated data centers.

SambaNova's rapid ascent from a startup to an $11 billion company in just months reflects a fundamental truth about AI infrastructure: the next trillion-dollar opportunity isn't in building bigger models, but in running them efficiently at scale. As enterprises deploy AI agents and autonomous systems across their operations, companies that can deliver fast, cheap inference will shape the future of enterprise AI.